Cell line-specific oxidative stress in cellular toxicity: A toxicogenomics-based comparison between liver and colon cell models.

Imbalance between high reactive oxygen species formation and antioxidant capacity in the colon and liver has been linked to increased cancer risk. However, knowledge about possible cell line-specific oxidative stress-mechanisms is limited. To explore this further, gene expression data from a human liver and colon cell line (HepG2/Caco-2), both exposed to menadione and H2O2 at six time points (0.5-1-2-4-8 and 24h) were compared in association with cell cycle distribution. In total, 3164 unique- and 1827 common genes were identified between HepG2 and Caco-2 cells. Despite the higher number of unique genes, most oxidative stress-related genes such as CAT, OGG1, NRF2, NF-κB, GCLC, HMOX1 and GSR were differentially expressed in both cell lines. However, cell-specific regulation of genes such as KEAP1 and GCLM, or of the EMT pathway, which are of pathophysiological importance, indicates that oxidative stress induces different transcriptional effects and outcomes in the two selected cell lines. In addition, expression levels and/or -direction of common genes were often different in HepG2 and Caco-2 cells, and this led to very diverse downstream effects as confirmed by correlating pathways to cell cycle changes. Altogether, this work contributes to obtaining a better molecular understanding of cell line-specific toxicity upon exposure to oxidative stress-inducing compounds.

Time series analysis of oxidative stress response patterns in HepG2: a toxicogenomics approach.

Oxidative stress plays an important role in chemically induced liver injury, however, our insight into molecular responses to different oxygen radicals is fragmentary. Since these cellular responses will differ over time, examining time-dependent changes in gene expression, and correlating these with markers for oxidative stress, may provide new insights into responses to oxidants. We used the human hepatoma cell line HepG2 to investigate the effects of oxidative stress on the transcriptome level by micro-arrays at seven time points (0.5, 1, 2, 4, 6, 8 and 24h) following exposure to the oxidants menadione, hydrogen peroxide and tert-butyl hydroperoxide including the effects on cell cycle and apoptosis by flow cytometry, protein carbonyl formation by spectrophotometry and oxidative DNA damage by FPG-comet. In total, 3429 genes were differentially expressed, including 136 genes that were significantly modified by all oxidants. Time-dependent biological pathway analysis showed that these genes were particularly involved in inflammatory responses, cell cycle processes and glutathione signaling. These responses were confirmed and supported by phenotypic anchoring to the different cellular endpoints. In addition, using an innovative temporal analysis we established an oxidative stress-related gene expression time cluster. Altogether, this study provides new insights in temporal oxidative stress mechanisms and demonstrates sequential cellular responses that may contribute to a better hazard identification and the mechanisms of toxicological responses in the liver induced by oxidative stress.

A transcriptomics-based in vitro assay for predicting chemical genotoxicity in vivo.

The lack of accurate in vitro assays for predicting in vivo toxicity of chemicals together with new legislations demanding replacement and reduction of animal testing has triggered the development of alternative methods. This study aimed at developing a transcriptomics-based in vitro prediction assay for in vivo genotoxicity. Transcriptomics changes induced in the human liver cell line HepG2 by 34 compounds after treatment for 12, 24, and 48 h were used for the selection of gene-sets that are capable of discriminating between in vivo genotoxins (GTX) and in vivo nongenotoxins (NGTX). By combining transcriptomics with publicly available results for these chemicals from standard in vitro genotoxicity studies, we developed several prediction models. These models were validated by using an additional set of 28 chemicals. The best prediction was achieved after stratification of chemicals according to results from the Ames bacterial gene mutation assay prior to transcriptomics evaluation after 24h of treatment. A total of 33 genes were selected for discriminating GTX from NGTX for Ames-positive chemicals and 22 for Ames-negative chemicals. Overall, this method resulted in 89% accuracy and 91% specificity, thereby clearly outperforming the standard in vitro test battery. Transcription factor network analysis revealed HNF3a, HNF4a, HNF6, androgen receptor, and SP1 as main factors regulating the expression of classifiers for Ames-positive chemicals. Thus, the classical bacterial gene mutation assay in combination with in vitro transcriptomics in HepG2 is proposed as an upgraded in vitro approach for predicting in vivo genotoxicity of chemicals holding a great promise for reducing animal experimentations on genotoxicity.

Comparison of phenotypic and transcriptomic effects of false-positive genotoxins, true genotoxins and non-genotoxins using HepG2 cells.

The conventional in vitro assays for genotoxicity assessment of chemicals are characterised by a high false-positive rate, thus failing to correctly predict their in vivo genotoxic effects. This study aimed to identify the cellular mechanisms induced by the false-positive genotoxins quercetin, 8-Hydroxyquinoline and 17-beta oestradiol in comparison to true genotoxins and non-genotoxins, by combining in vitro phenotypic parameters with transcriptomics data from HepG2 cells. The effects of these compounds on the phosphorylation of H2AX, cell cycle distribution and whole genome gene expression following treatment for 12, 24 and 48 h were compared with the effects of true genotoxins [benzo[a]pyrene and aflatoxin B1] and non-genotoxins (2,3,7,8-tetrachlorodibenzodioxin, cyclosporin A and ampicillin C). Quercetin induced similar phenotypic effects as true genotoxins and to some extent similar gene expression alterations. Different gene expression changes were also observed, including the up-regulation of DNA repair-related genes. 8-Hydroxyquinoline and 17-beta oestradiol showed no similarities to the true genotoxins at both the phenotypic and the transcriptomic level. In a classification approach, classifiers were selected to discriminate between genotoxins and non-genotoxins. Subsequent analysis for the false-positive compounds showed quercetin to be predicted as genotoxic and 8-hydroxyquinoline and 17-beta oestradiol as non-genotoxic. Our results support that transcriptomics analysis of compound effects in HepG2 leads to similar results with phenotypic analysis and provides additional mechanistic information. Therefore, combined evaluation of gene expression alterations and relevant functional end points using HepG2 cells may contribute to the better understanding of modes-of-action of chemicals and the correct evaluation of their genotoxic properties.

DNA copy number status is a powerful predictor of poor survival in endocrine pancreatic tumor patients.

The clinical behavior of endocrine pancreatic tumors (EPTs) is difficult to predict in the absence of metastases or invasion to adjacent organs. Several markers have been indicated as potential predictors of metastatic disease, such as tumor size > or =2 cm, Ki67 proliferative index > or =2%, cytokeratin (CK) 19 status, and recently in insulinomas, chromosomal instability (CIN). The goal of this study was to evaluate the value of these markers, and in particular of the CIN, to predict tumor recurrence or progression and tumor-specific death, using a series of 47 insulinomas and 24 non-insulinoma EPTs. From these EPT cases, a genomic profile has been generated and follow-up data have been obtained. The proliferative index has been determined in 68 tumors and a CK19 expression pattern in 50 tumors. Results are statistically analyzed using Kaplan-Meier plots and the log-rank statistic. General CIN, as well as specific chromosomal alterations such as 3p and 6q loss and 12q gain, turned out to be the most powerful indicators for poor tumor-free survival (P< or =0.0004) and tumor-specific death (P< or =0.0113) in insulinomas. The CIN, chromosome 7q gain, and a proliferative index > or =2% were reliable in predicting a poor tumor-free survival in non-insulinoma EPTs (P< or =0.0181, whereas CK19 expression was the most optimal predictor of tumor-specific death in these tumors. In conclusion, DNA copy number status is the most sensitive and efficient marker of adverse clinical outcome in insulinomas and of potential interest in non-insulinoma EPTs. As a consequence, this marker should be considered as a prognosticator to improve clinical diagnosis, most practically as a simple multi-target test.

Molecular parameters associated with insulinoma progression: chromosomal instability versus p53 and CK19 status.

Insulinomas represent the predominant syndromic subtype of endocrine pancreatic tumors (EPTs). Their metastatic potential cannot be predicted reliably using histopathological criteria. In the past few years, several attempts have been made to identify prognostic markers, among them TP53 mutations and immunostaining of p53 and recently cytokeratin 19 (CK19). In a previous study using conventional comparative genomic hybridization (CGH) we have shown that chromosomal instability (CIN) is associated with metastatic disease in insulinomas. It was our aim to evaluate these potential parameters in a single study. For the determination of CIN, we applied CGH to microarrays because it allows a high-resolution detection of DNA copy number changes in comparison with conventional CGH as well as the analysis of chromosomal regions close to the centromeres and telomeres, and at 1pter-->p32, 16p, 19 and 22. These regions are usually excluded from conventional CGH analysis, because they may show DNA gains in negative control hybridizations. Array CGH analysis of 30 insulinomas (15 tumors of benign, eight tumors of uncertain and seven tumors of malignant behavior) revealed that >or=20 chromosomal alterations and >or=6 telomeric losses were the best predictors of malignant progression. A subset of 22 insulinomas was further investigated for TP53 exon 5-8 gene mutations, and p53 and CK19 expression. Only one malignant tumor was shown to harbor an arginine 273 serine mutation and immunopositivity for p53. CK19 immunopositivity was detected in three malignant tumors and one tumor with uncertain behavior. In conclusion, our results indicate that CIN as well as telomeric loss are very powerful indicators for malignant progression in sporadic insulinomas. Our data do not support a critical role for p53 and CK19 as molecular parameters for this purpose.

Novel candidate tumour suppressor gene loci on chromosomes 11q23-24 and 22q13 involved in human insulinoma tumourigenesis.

Insulinomas represent the predominant syndromic subtype of endocrine pancreatic tumours. Previous molecular studies have shown that gain of chromosome 9q rather than MEN1 gene mutation is an important early event in tumour development and that chromosomal instability is associated with metastatic disease. In order to identify new gene loci and to define further the critical genetic events in insulinoma tumourigenesis, 27 insulinomas were investigated by array-based comparative genomic hybridization (array CGH) on 3.7 k genomic BAC arrays (resolution < or =1 Mb). Fluorescence in situ hybridization was used to validate alterations in a subset of tumours. Array CGH most frequently detected loss of chromosomes 11q and 22q and gains of chromosome 9q. The chromosomal regions of interest (CRI) included 11q24.1 (56%), 22q13.1 (67%), 22q13.31 (56%), and 9q32 (63%). Evaluation of the simultaneous occurrence of these aberrations in the individual tumours revealed that gain of 9q32 and loss of 22q13.1 are early genetic events in insulinomas, occurring independently of the other alterations. In tumours with increased genomic complexity, these alterations were often detected simultaneously, occurring in the same tumour cells. Losses of 11q24.1 and 22q13.31 were also associated with these more advanced tumour cases. The CRIs identified most likely harbour crucial candidate genes important in insulinoma tumourigenesis.

Chromosomal instability predicts metastatic disease in patients with insulinomas.

Endocrine pancreatic tumors (EPTs) comprise a highly heterogeneous group of tumors with different clinical behavior and genetic makeup. Insulinomas represent the predominant syndromic subtype of EPTs. The metastatic potential of insulinomas can frequently not be predicted using histopathological criteria, and also molecular markers indicating malignant progression are unreliable because of the small number of cases per subtype studied so far. For the identification of reliable indicators of metastatic disease, we investigated 62 sporadic insulinomas (44 benign and 18 tumors with metastases) by means of comparative genomic hybridization (CGH). In addition, the role of MEN1 (multiple endocrine neoplasia type 1) gene mutations was determined to assess specific chromosomal alterations associated with dysfunction of this endocrine tumor-related tumor suppressor gene. Only one case with a somatic MEN1 mutation was identified (1527del7bp), indicating that the MEN1 gene plays a minor pathogenic role in sporadic insulinomas. CGH analysis revealed that the total number of aberrations per tumor differs strongly between the benign and the malignant group (4.2 vs 14.1; P<0.0001). Furthermore, chromosome 9q gain was found to be the most frequent aberration in both benign and malignant insulinomas, whereas chromosome 6q losses and 12q, 14q and 17pq gains are strongly associated with metastatic disease. Our study shows that chromosomal instability, as defined by > or =5 gains together with > or =5 losses, or total number of gains and losses > or =8, rather than parameters such as tumor size and proliferation index, is the most powerful indicator for the development of metastatic disease in patients with sporadic insulinoma.